• Gloria C Chiang, Naveen Galla, Richard Ferraro, and Ilhami Kovanlikaya.
• Department of Radiology, Division of Neuroradiology, Weill Cornell Medical College, NewYork-Presbyterian Hospital, New York, NY.
• J Neuroimaging. 2017 Mar 1; 27 (2): 243-247.

Background And PurposePatients with glioblastoma multiforme (GBM) face a dismal prognosis, with an average survival of 6-7 months after recurrence. There remains no consensus for managing these patients due to the heterogeneity of these tumors. Imaging may affect treatment decisions by helping to stratify patient prognosis. The purpose of this analysis was to evaluate the added utility of 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) over magnetic resonance (MR) imaging metrics in predicting survival.MethodsForty-four consecutive patients who underwent FDG-PET for first recurrence of GBM were included in this analysis. Tumor sizes, using cross products, and volumes on FDG-PET and MRI, maximum standardized uptake value (SUV), minimum apparent diffusion coefficient (ADC) value, presence of satellite lesions, presence of multifocal lesions, and presence of bilateral tumor were considered as prognostic variables. Survival was assessed using Cox hazard and logistic regression models based on the time interval between the PET scan and the patient's date of death.ResultsTumor volumes on FDG-PET (P = .046), tumor cross products on FDG-PET (P = .017), and tumor cross products on MRI (P = .031) were significant prognostic variables, adjusting for the extent of the initial resection. Enhancing tumor volume, tumor cross product on a T2-weighted MRI sequence, maximum SUV on FDG-PET, minimum ADC value, presence of satellites, multifocality, and bilaterality were not prognostic (P > .5). Prognostic accuracy of predicting short survival increased from 58% with tumor cross product on MRI alone to 74% after including tumor cross product from PET.ConclusionsTumor size on FDG-PET adds prognostic information to enhancing tumor size on MRI at first suspected recurrence of GBM.Copyright © 2016 by the American Society of Neuroimaging.

21 keywords...

hide…